web analytics 101
DESCRIPTION
The story and thought process for analytics. Before you get down to using analytics, understand the 'hows' and 'whys'.TRANSCRIPT
Web Analytics 101
Saurabh Pandey
What is Web Analytics?
• Web analytics is the measurement, collection, analysis and reporting of internet data for the purposes of understanding and optimizing web usage.
Lay-Man’s Term
• What is the key metrics of success for my website?
• How am I faring on them?
Why Web Analytics?
Analytics Timeline
Web Analytics Vs. Log Files
• Major Challenges in Log Files Data
– Browser Caching leading to data inaccuracy
– Uniqueness of visitors not accurate- largely due to usage of proxy servers by the ISPs the actual IP address are never reflected.
– Due to the same proxy caching the accuracy of traffic volume may also be questionable
Web analytics use Page tagging to tag
Each and every page of the website.
Due to this and the use of cookies to identify
Unique visitors , the data is much more
Accurate.
It’s easier to access , view and report data
However the downside:
1. Data is generally on cloud, and not on your
servers.
1. About 31% people delete cookies in a
month- this puts question mark on data
accuracy
Types of Web Analytics
• Off site
• Onsite
• Onsite more popular
• Beginning to use a hybrid-mix
Off-Site Analytics Services
• Work on a sample of Panel users (e.g. compete works with a sample of 2m panel users)
• No code (cookies) inserted into the site/browser
• Less accurate
• Good for measuring trends of competition
Offsite-Alexa,Compete, SocialMention
More useful to visualize trends and buzz
Around competition and your own site.
No need to be the admin, or insert codes
On-Site Web Analytics Services
• Lends deep insight into
– how users find and use our site.
– how to make the product (website) more efficient
– how to make marketing more efficient
Various Brands
• Webtrends-Paid
• Omniture/Site-Catalyst-Paid
• Coremetrics-Paid
• Google Analytics-Free
The paid analytics systems can go upto Rs. 50 odd lacs (USD 100K or more) a year.
Paid analytics services have many advanced segmentation and tracking features,
Coupled with data privacy and hosting facility.
They have extensive and regular training & support systems in place.
Google Analytics, though, serves the purpose for all SMBs and many large enterprises,
Data resides on the cloud. Lacks full-blown support system
How does GA work
Understanding KPIs & Metrics
Not everything that can be counted counts!
Albert Einstien
KPIs/Metrics
• Metrics are
– Count: e.g. No. of visitors OR
– Ratios: e.g. Conversion rate (Buyers/Visitors)
• KPIs are also metrics but we use them to see how well we fare against our business objectives!
KPIs/Metrics
• If I run Flipkart. – My Business Objective could be to sell more books
– My KPI would be: Average Order Size &Total Units Sold.
– My Metrics would be: Conversion rate and visitors
– KPIs are clearly tied to an objective.
– Business objectives>Goals>KPIs>Metrics
What’s Next? Targets
• So, now we know – Our Objectives – Our KPIs and – Our Goals
– But we still need to know whether we are doing great or not?
– We need to create Targets for each Metric. Better still, we
should have 1 target for each KPI.
– E.g. My Average Order Value Target at Flipkart should be Rs. 250/-
Dimensions of a visitor/transaction!
• Attributes of the visitor
– Which Keyword ?
– What landing page?
– Which part of the country?
– What browser?
– What time?
– Is this the 3rd visit?
– How many pages viewed?
Remember STAR
◦ Source (referring URLs, Geography etc.)
◦ Technical (browser, screen resolution, ISP etc.)
◦ Activity Performed (landing page, content path, time spent, produts they bought, downloaded something etc.)
◦ Recency (when was the last time they visited)
Segments are Important…
• In making sense of data and getting real insights.
• To be frank- individually almost all data is useless. • But if you combine metrics with dimension (segments) it starts
giving you insights.
• For e.g. No. of visits- is useless • But look at this:
– No. of Visits from Delhi or – no. of visits that lasted more than 1 minute – No. of visits from Google that stayed less than 1 minute.
– Are all much more meaningful now.
Segmentation in Google Analytics
We can see a trend of bounced visits from ‘facebook’. We can also see
That almost 80% of these bounces are from new visits (first time)
• Let’s Create an Analytics Framework now!
Web Analytics Framework
Analytics-KPIs (Process Centric Understanding)
PageViews
Bounce
Time Spent
Content
Visits Unique Visitors
Source Of Traffic
Technical
Conversion % No. of
Transactions
Average Order Value
Products Sold
Engagement & Behavior Acquisition & Reach Transaction
Let’s Understand Some of the Major KPIs
Hits- Obsolete metric. Each time a server is requested for a file, it’s considered a hit
Hence flawed. If a page has 10 images, then the server will show 10-11 hits for just one page view
Hence need for more robust and realistic metrics of Key Performance Indicators
Google Analytics Dashboard
Visits
• Whenever someone comes and views the web-pages, it’s counted as 1 visit.
• If 100 pages are viewed without exiting the website, the visit is still counted as 1.
• The visit ends when the user exits from the website. Or
• When the session expired. 30 Minutes
Page View
• An instance of a webpage loaded by the browser.
Pages per visit
• Number of pages viewed in a single visit.
• If a user views 2 pages in a single visit then Pages/Visit = 2
Visitor
• Visit ≠Visitor
• A user that visits your site.
• Any future session by the same visitor will be counted as 2nd visit but not as a 2nd visitor.
• 30 minute session
Bounce
• Number of single page visits.
• Entry and exit from the same page
Average Time on Site
• Total amount of time spent on the website
• Divided by the total number of visitors
Content- Landing Page & Navigation Flow
We can see the top pages in terms of views or the navigation flow and
other interesting metrics around content engagement
Snapshot-Analytics Dashboard
Understand the traffic sources
• Direct
• Referral
• Search
Direct
• All traffic (visits/visitors) that come to the website by:
– Typing in the URL in the browser
– Clicking on a link in Bookmarks
Referral
• All visits that come by clicking on a link somewhere.
• For e.g. If you are reading an article on a Rediff.com, and you found a link to www.atomthought.com which you clicked then:
• This is a referral visit for atomthought.com
• Rediff.com becomes the referrer site.
Organic Search
• All visits by clicking a link in the organic search results page (not paid search results)
Geo-Where on earth the traffic is coming from?
Track Goals-Funnel Visualization
Understand entry and exit points, wastage in the funnel, and get an
overview of where to fix things
What Kind of Goals can be set?
• URL destination
• Time Spent
• Pages/Visit
Goal Setting
You have 4 Goal Sets, each with 5 Goals (You can set a total of 20 goals)> Add Goal
Let’s come to Optimization
• Website Objectives>Goals>KPIs> Test & Optimize
Web-Optimization
• A/B Testing &
• Multivariate Testing
• A/B testing is the best to start with. Absolutely great for SMBs.
A/B Testing
• Most often people don’t come to the ‘Home Page’
• People come rather to a specific landing page
• A good landing page can make a lot of difference to conversions and hence revenues
• So, how do we know which landing page to go ahead with?
A/B Testing With Google Website Optimizer
• Choose the page you want to test • Decide the testing parameters
– Pick a few elements (ideally 1, and worse situation 2-3) like:
• Headline • Price • Product Image • Offer • Font etc.
(A/B testing is a convenient and easy way to decipher which landing page is working for you. Changing few parameters will give meaningful results)
• Log into Google Website Optmizer (google.com/websiteoptimzer)
– Google calls these tests as ‘experiments’
A/B Testing With Google Website Optimizer
LP1
LP2
ThankYou
A/B Testing With Google Website Optimizer
• Install Javascript Code: Now the website optimizer knows which pages to track, but in order for it to show these pages randomly to all visitors, we need to put a snippet of code in the Control Page, Variation Page and Confirmation Page. (just after /header in your html page for control and before /body in others)
• After doing this just press ‘Validate’ in Google Website Optmizer.
A/B Testing With Google Website Optimizer
Reports
This is a multivariate report. As you can see, Combination 11 looks to be the best.
How to create Tracking URLS and start tracking
Setting up
Go to http://analytics.google.com Signin with your gmail id; or sign up
for a new account
Setting up
Agree to the t&c
Get the Code
Where to put the code?
• Just before the end of closing tag(/body)
• Each page of your website
• If it’s a wordpress hosted site- go to footer.php (in your themes tab)
What if I have a blogspot blog?
Conclusion
• There’s more to analytics than what we just experienced.
• But start simple. Scale gradually. Focus on Core Metrics
• You should be able to segmentize your metrics.
• Use both offline and online analytics
Reach Me @
• Twitter: pandeysaurabh • Blog: www.atomthought.com
• Get more presentations on Social Media & Digital
Marketing @ – Slideshare: http://www.slideshare.net/sycamore
• Thanks!